Archives ofDisease in Childhood 1990; 65: 1060-1062
Outcome after acute osteomyelitis in preterm infants J B Williamson, C S B Galasko, M J Robinson
Abstract Eight cases of skeletal infection in preterm infants were studied. All the infants were systemically unwell, with polymorpholeucocytosis. Diagnosis was by blood culture, and any radiographic changes were apparent at the time of presentation. Infection was often multifocal, with sites around the knee being most commonly affected. Staphylococcus aureus was the pathogen isolated in six of the eight cases; in these treatment with fusidic acid was effective and well tolerated, even at doses that were less than the recommended therapeutic minimum. Even with prompt diagnosis and aggressive treatment orthopaedic sequelae are common.
Acute osteomyelitis is an uncommon but important neonatal infection. In this paper we report the methods of treatment and results in eight preterm patients with acute osteomyelitis. Patients and methods Eight preterm infants with skeletal infections were admitted to our neonatal intensive care unit between 1982 and 1988; their salient clinical features are summarised in the table. All eight children were preterm, being born between 25 and 33 (mean 29-4) weeks' gestation. All were admitted to the neonatal unit for treatment of the consequences of prematurity. Four of the eight had umbilical arterial lines in place at the time of the development of the septic episode. In each patient sepsis was suspected on clinical grounds before the presentation of any localising signs. All eight patients had considerably increased white cell counts. In all eight the clinical presence of infection prompted the performance of a 'septic screen'. Organisms were isolated from the primary blood cultures in each patient and in addition three patients grew organisms from other sites-the umbilical stump, cerebrospinal fluid, and an infected intravenous cannula. The pathogens isolated were Staphylococcus aureus in six of the eight patients, and group G , Hope Hospital, haemolytic streptococcus and Candida albicans Salford M6 8HD, in the two. remaining of Manchester University There was a delay (mean 2-9 days, range 0Department of Orthopaedic Surgery 8)-especially early in the series-between preJ B Williamson sentation with systemic signs and the definitive C S B Galasko Regional Neonatal Unit diagnosis of skeletal sepsis. In the most recent four patients the definitive diagnosis was made M J Robinson on the day of presentation. Correspondence to: Mr Williamson. Five of the eight patients were treated by Accepted 24 April 1990 some form of drainage procedure (aspiration,
n= 3; open surgical drainage, n=2). All patients were treated with sodium fusidate, 20 mg/kg! day, initially intravenously (for a mean of 11 days, range 7-21), and orally when they were tolerating oral feeds (for a mean of 22 days, range 21-27). Treatment lasted for a mean of 32 days (range 28-42). An additional antistaphylococcal antibiotic was used when appropriate. Initially serum concentrations of fusidic acid were measured by high performance liquid chromatography. Despite clinical improvement in all eight patients, therapeutic concentrations of fusidic acid were not achieved with either oral or parenteral treatment. Serum bactericidal titres were also below the recommended therapeutic minimum. The duration of treatment was determined by each infant's clinical progress, and in one patient it continued after discharge from hospital. Six patients were reviewed in the follow up clinic. The recent progress of the two nonattenders was obtained from their general practitioners. Results The patients had been followed up for between 2 years 2 months and 6 years since their infections. Two of the eight children failed to attend the specially arranged follow up clinic, but had normal musculoskeletal examinations when last seen in the orthopaedic clinic at 3 5 years of age. Skeletal sepsis seemed to have been abolished by treatment in all patients. At the time of follow up no patient had developed chronic
osteomyelitis. In six of the eight patients the radiographic appearance was diagnostic of skeletal infection. The presence of radiographic changes did not seem to be related to the length of time that sepsis had been present, as there were changes on the radiographs of patients taken on the day that they had presented with systemic upsets. Two patients had normal radiographs throughout, and in two (one of whom had a diagnostic radiograph) 99mTc labelled diphosphonate scintigraphy was carried out; both scintigrams were normal. Long term orthopaedic sequelae, such as discrepancies in limb length, angular deformity, and joint stiffness developed in five of the eight patients. Progressive discrepancy in limb length was the commonest long term problem (n= 5), with joint destruction and axial deformity each occurring in one patient. One patient (case 7) has had his deformity corrected surgically. Limb lengthening is planned for the same patient in the near future.
Outcome after acute osteomyelitis in preterm infants
Clinical details of eight neonates with acute osteomyelitis Case No
Birth weight Gestation (g) (weeks)
Age at presentation
Day 5 systemic upset: blood culture Day 11 left distal femur
Apnoea S aureus in blood and Swollen knee aspirate White cell count 22 00Ox 109/l Day 13 right distal femur Radiograph normal Day 19 left distal tibia Scintigram normal Day 10 systemic upset: White cell count S aureus in blood and blood culture 25 OOOx 109/l aspirate Day 13 left proximal Radiograph hip: femur dislocation, gross osteomyelitis S aureus in blood and Day 7 systemic upset: White cell count blood culture 45 OOOx 109/l aspirate Day 9 right distal femur Radiograph: osteomyelitis S aureus in blood and White cell count Day 7 systemic upset: 29 000X 109/f cerebrospinal fluid cerebrospinal fluid and blood culture Radiograph: Day 19 right distal femur osteomyelitis Day 21 right proximal Scintigram normal femur Day 21 left proximal tibia Hot swollen knee S aureus in blood and Day 15 systemic upset: blood culture White cell count aspirate Day 15 right distal femur 29 00Ox 109/l Day 17 left proximal tibia Radiograph normal Day 7 systemic upset: Abdominal distension S aureus in blood and blood culture White cell count aspirate left proximal femur 30 OOOx 109/1 Radiograph: osteomyelitis C albicans in blood White cell count Day 19 umbilical line 25 00Ox 109/l and catheter tip Day 49 left distal femur Radiograph: osteomyelitis Day 10 skin abscess White cell count ,l haemolytic Day 16 right distal femur 23 OOOx 109/l streptococcus in pus from skin abscess, Day 16 right proximal Radiograph: femur and from blood osteomyelitis
Aspiration; fusidic acid and erythromicin
Full movements; 2 cm shortening at 6 years of age
Surgical drainage; fusidic Decreased movement; 3 cm shortening at acid and erythromicin 2-2 years of age
Aspiration; fusidic acid and flucloxacillin
No sequelae at 3 years of age
No drainage; flucloxaciUin and
No sequelae at 3 5 years of age
followed by fusidic acid Surgical drainage; fusidic acid and erythromicin
No sequelae at 3 5 years of age
Aspiration; fusidic acid and flucloxacillin
Full movement; 1 cm shortening at 3 5 years of age
30' Flexion deformity at knee; 8 cm shortening at 6 5 years of age No drainage; fusidic acid No sequelae at 3 5 years of age and penicillin No drainage; amphotericin and cefuroxime
Case 7 had a candida infection of his umbilical line on day 16 that persisted. He became 'septic' on day 49 with a candida septicaemia and osteomyelitis. Case 8 had a skin infection on day 10, but did not become systemically unwell until day 16, when she had localising signs of skeletal sepsis.
The orthopaedic outcome was not related to operation made good recoveries from their systhe severity of initial infection, delay in making temic upsets once treatment with antibiotics a definitive diagnosis, or the performance of was started. Other authors have reported the occurrence surgical drainage. of bone and joint infection distal and ipsilateral to the tip of umbilical artery catheters, espeDiscussion cially if vascular spasm accompanied its Osteomyelitis is an uncommon but important introduction.6 It was suggested that osteomyelineonatal infection with a recognised morbidity tis associated with umbilical catheterisation may and mortality.' The definitive treatment and its be avoided by adherence to a protocol incorporduration has not yet been defined. ating strict asepsis and immediate removal of Osteomyelitis in well nourished infants born the catheter if there was vascular compromise. at full term seems to cause little in the way of In our patients who had umbilical arterial cathesystemic upset,2 3 but constitutional symptoms terisation, the catheter tip was at L4/5, and were much more of a feature in a series of there was no suggestion of vascular insuffiNigerian infants born at full term. 1 Two of the ciency. The distribution of the affected joints four patients with osteomyelitis associated with was not significantly different from that in the umbilical catheter sepsis reported by Brill et al patients without umbilical catheters, and we do were systemically unwell,4 but two had no gen- not think that umbilical arterial catheterisation eralised features of sepsis. Only one of their was an important factor in the development of babies who had systemic symptoms was pre- sepsis. The presence of radiographic signs at presenterm. All eight of our patients were preterm and all had systemic symptoms. We suggest that tation in neonates has not been noted previously osteomyelitis in infants who are immunocom- to our knowledge, and may suggest that the promised because of malnourishment or pre- infection had been present subclinically before maturity is much more likely to be accompanied presentation. The limited value of skeletal scinby septicaemia and systemic upset. Multifocal tigraphy was commented on by Ash and disease occurred in four of our patients and this Gilday.7 The effectiveness and lack of serious side accords with the experience of other authors.5 Blood cultures from all our patients grew effects of fusidic acid in the treatment of stapathogens, and were helpful in deciding defmi- phylococcal neonatal osteomyelitis is well tive antimicrobial treatment. When drainage documented.8 9 The use of a second antistaphyprocedures were carried out they served both to lococcal drug is essential to prevent the emergdecompress and to aid diagnosis. Identical ence of resistant strains.10 The fact that adult organisms were grown from the aspirate and the therapeutic concentrations or recommended blood cultures. All patients treated without bactericidal concentrations were never achieved
Williamson, Galasko, Robinson
despite considerable clinical improvement is interesting, and illustrates the difference between the in vivo and in vitro effects of drugs. It is a matter of orthopaedic tradition and dogma that long courses of antibiotic (usually six weeks) are used for skeletal infection. This does not seem to be necessary in preterm infants. The distinction between septic arthritis and osteomyelitis is often difficult in infants. The epiphyseal circulation is derived from the same periarticular anastamoses as the metaphyseal circulation and this common circulation provides a pathway for the transphyseal spread of infection. It would therefore seem likely that the primary skeletal infection is metaphyseal osteomyelitis, but a secondary septic arthritis soon develops. The growth plate is surrounded by infected tissue and its blood supply, which is almost exclusively from the metaphyseal circulation, is compromised. Most of the orthopaedic complications were the results of physeal damage by the septic process, with progressive discrepancy in limb length being caused by symmetrical physeal damage, and axial deformity being caused by asymmetrical physeal damage. CONCLUSIONS
In preterm infants the clinical picture of acute osteomyelitis is different from that in babies
born at full term and in older children, in that systemic upset and pronounced polymorpholeucocytosis are invariably present. Any radiographic changes were apparent at the time of presentation. It is important to keep the possibility of skeletal sepsis in mind, and any child with a clinical diagnosis of sepsis, especially staphylococcal septicaemia without an obvious focus, should have the hips and knees carefully examined and radiographed. Treatment with sodium fusidate and a second antistaphylococcal drug, initially given intravenously, achieved satisfactory antimicrobial effect, although bony sequelae were common. 1 Omene JA, Odita JC, Okolo AA. Neonatal osteomyelitis in Nigerian infants. Pediatr Radiol 1984;14:318-22. 2 Fox L, Sprunt K. Neonatal osteomyelitis. Pediatrics 1978;62: 535-42. 3 Weissberg EG, Smith AL, Smith DH. Clinical features of acute osteomyelitis. Pediatrics 1974;53:505-10. 4 Brill PW, Winchester P, Krauss AN, Symchych P. Osteomyelitis in a neonatal unit. Radiology 1979;131:83-7. 5 Mok PM, Reilly BJ, Ash JM. Osteomyelitis in the neonate. Radiology 1982;145:677-82. 6 Lim MO, Gresham EL, Franken EA, Leake RD. Osteomyelitis as a complication of umbilical artery catheterization. Am J Dis Child 1977;131:142-4. 7 Ash JM, Gilday DL. The futility of bone scanning in neonatal osteomyelitis: concise communication. J Nucl Med 1980; 21:417-20. 8 Bergdhal S, Elinder D, Eriksson M. Treatment of neonatal osteomyelitis with cloxacillin in combination with fusidic acid. Scan J Infect Dis 1981;13:281-2. 9 Liddy N. Intravenous fusidic acid in the newborn. Lancet 1973;i:621. 10 Mollan RAB, Piggot J. Acute osteomyelitis in children. J Bone joint Surg 1977;54B:2-7.